بهينه سازي نرخ حفاري (ROP) به كمك كوپل شبكه عصبي مصنوعي با الگوريتم ژنتيك

Optimization of Rate of Penetration (ROP) using Artificial Neural Network coupled with Genetic Algorithm

A new method which is able to predict and optimize the rate of penetration (ROP), is presented in this article using Artificial Neural Network coupled with Genetic Algorithm. ROP is an important parameter for drilling operation. An accurate prediction and optimization of ROP may immensely affect other parameters, for example it can lead to longer bit lifetime or decreased drilling costs. Also, there are many parameters which affect the ROP and the predicted ROP. Different methods are presented for ROP prediction among which, Bourgoyne and Young (1974) method is mostly used to predict the ROP however this method cannot predict ROP accurately in some cases. In this article, an Artificial Neural Network model is used with Genetic Algorithm as its training function, and is coded through MATLAB software. Malaysia Kinabalu oil field data and Iran Ahvaz oil field data are used to develop and evaluate the created model. According to the results which are achieved using few data sets of both fields, it can be concluded that more accurate predictions could be resulted from larger data sets.

A new method which is able to predict and optimize the rate of penetration (ROP), is presented in this article using Artificial Neural Network coupled with Genetic Algorithm. ROP is an important parameter for drilling operation. An accurate prediction and optimization of ROP may immensely affect other parameters, for example it can lead to longer bit lifetime or decreased drilling costs. Also, there are many parameters which affect the ROP and the predicted ROP. Different methods are presented for ROP prediction among which, Bourgoyne and Young (1974) method is mostly used to predict the ROP however this method cannot predict ROP accurately in some cases. In this article, an Artificial Neural Network model is used with Genetic Algorithm as its training function, and is coded through MATLAB software. Malaysia Kinabalu oil field data and Iran Ahvaz oil field data are used to develop and evaluate the created model. According to the results which are achieved using few data sets of both fields, it can be concluded that more accurate predictions could be resulted from larger data sets.